Unit Root Tests Based on Instrumental Variables Estimation
基于工具变量估计提出新的单位根检验方法，在移动平均误差下渐近有效，有限样本下准确，且比同类检验更有效。
This paper develops new tests of the unit root hypothesis based on instrumental variables estimation. The tests are asymptotically valid in the presence of moving average errors and they are quite accurate in finite samples. They are more powerful against stationary alternatives than other tests that are equally accurate under the null. Copyright 1994 by Economics Department of the University of Pennsylvania and the Osaka University Institute of Social and Economic Research Association.